$500k of Azure credit for YC startups(blog.ycombinator.com) |
$500k of Azure credit for YC startups(blog.ycombinator.com) |
http://www.networkworld.com/article/2866950/cloud-computing/...
That's can't be the only reason to choose (or ignore) a particular platform, because:
- 100% is not commercially feasible
- with rare exceptions, a little downtime won't kill you
Azure had 39.77 hours of downtime, whilst Rack space had 7.52 and the winner (of the selected cloud providers was AWS with 2.41 hours.
I'm not sure how those figures were calculated though. These cloud providers are multifaceted. I'm not sure which parts if each's services were analysed.
Also, it is worth considering that one serious problem can distort these averages. AWS has a much more stable long term history.
In my opinion the competition is a good thing and Microsoft seem happy to support whatever framework or OS you want. Their node.js support is supposed to be pretty good.
The marketplace also enables you to get things like MongoDb clusters, etc which is competition for SQL Server. They seem to be interested in supporting customer interests rather than the old fashioned MS tech or die philosophy from back in the day.
Factors for downtime were an update bug in November and weather. The bug is very unlikely to happen again. Outages due to weather really don't tell you much at all other than that they were unlucky. Overall, I would say the downtime numbers are not very useful.
http://www.computerweekly.com/news/2240238379/Microsoft-Azur...
E.g., boot up a bitcoin mining cluster (though this particular idea would obviously yield low return since it would be running on off the shelf hardware)
This would only benefit YC startups who already use Azure. (Unless the switching cost from AWS/Rackspace/Google is low enough nowadays.)
And for those that have a proof-of-concept running on some other provider(s) (AWS/Rackspace/Google), it would be a huge boon to be able to migrate and scale using Azure and the $500k credit.
Would definitely be harder if you have a more complex infrastructure - but in most cases entirely worth it.
I wondered why my company after 18 months of effort to get going on AWS with certain apps has had a shift to get on Azure 'without compromise' in 2 months.
I doubt these sort of offers are only going to start ups, Microsoft is taking on web services hard core.
Not saying you did not have a bad experience, just offering mine. :)
The PaaS front is of particular interest. For example, it's not immediately evident that Heroku even offers anything similar, beyond a small free tier available to anyone.
Professionally, for simplicity I'd choose Rackspace or Google. For Scalability and price I'd choose Amazon. For personal projects I love Digital Ocean.
I dislike Azure's control panel and VM setup, and find their VMs slow. I'd only really consider them if I planned to implement a project that required leveraging the entirety of the Microsoft technology stack.
I can't see technology companies embracing Virtual Windows Server RTs and MSSQL Servers. Startups enjoy technologies that are easy to manage and scale with smaller teams. Enterprise level companies still need inhouse dedicated hardware for security.
Azure is kind of an oddball in the virtual hosting space.
The oldest tactic in the Proprietary Platform Handbook is giving away free stuff to hook people early on.
In this case Azure can run Linux VMs, so I wouldn't worry much about using them for that. Linux lacks lock in.
Does anyone actually trust cloud services in Five-eyes jurisdictions anymore? Is there a company in those jurisdictions that people could use that is more dangerous to trust than Microsoft?
Is this sentence constructed poorly or did a private comment slip in?
Not sure how long we will last though...
And with Win Server, I run ASP.NET and PHP and Node, it works nice. Db (pg) is on linux.
I did some of my own quick benchmarks here http://browser.primatelabs.com/user/82525
For comparison, on-demand prices are (Linux / Windows):-
* AWS c3.xlarge - $0.239 / $0.376
* AWS c4.xlarge - $0.264 / $0.430
* Azure Standard A3 - $0.24 / $0.324
* Azure Standard D3 - $0.46 / $0.732
I also did some quick disk benchmarks https://drive.google.com/folderview?id=0B096UguGrNHAfmZjM1VQ...
My tests are very limited I realise, but it looks to me like the AWS boxes represent better value for the money there, even in the on-demand range. Azure used to offer 6-month and 12-month plans but they removed those. Now you need an enterprise agreement to get anything other than published rates, and you need a big spend to qualify for that AFAICT. AWS reserved instances therefore make their instance rates even more competitive, if those make sense for you.
I sent all this to the company trying to push Azure to us hoping they would have better data to refute my simple benchmarks, but they didn't really have any coherent response beyond talking about the "other advantages" of Azure.
I also looked at:-
* https://cloudharmony.com/reports/editions/state-of-the-cloud...
* http://windowsitpro.com/cloud/microsofts-unwanted-win-cloud-...
Beyond that I looked at the excellent data at https://cloudharmony.com/cloudscores (great site by the way). Their service is not live yet, so I scraped the data into Google Sheets and compared it there. AWS seems to beat Azure on most cases.
You can see my sheet here:- https://docs.google.com/spreadsheets/d/19jltGybjXN3-PnKXAwQX...
(Data entirely owned by Cloud Harmony). Sorry, I can't find a way to let viewers change the filter on the data.
As far as "other advantages go":-
* If you use Win Server 2012 + HyperV in house then there's a good reason to use Azure because of tight integration.
* If you're using SQL Server it's potentially a good fit too. (MySQL on Azure is recommended via a DB as a service they offer from a 3rd party which seems very pricey).
* If you're starting out and are using Visual Studio tooling then the tight integration with Azure Websites, Web Roles and Worker Roles would be pretty nice since the build/test/deploy process would be pretty slick. Re-engineering an existing system to fit the PaaS could be hard, depending on your architecture (it would be for us).
AWS might be slightly less, but I've found it more complicated to determine pricing. Plus AWS locks you in to specific instances (more or less) whereas Google just automatically applies it, nice and easy.
Of course I'm nervous about using Google for anything, but they have a very impressive offering.
Also, do competing services have anything similar to MS's machine learning services?
I've found the network drives to be oddly slow, but the machines otherwise to perform well. So they can be performant if you can run I/O-intensive stuff off the ephemeral local storage, especially on one of the SSD instances.
You are just hating.
Free Office365 and CloudFlare are probably the best part of this announcement, since they are "easily migrate-able" by most startups. Migrating from DO/AWS/Linode isn't fun and productive. Even though they are negligible (in terms of money and value) they are the easiest "sticky" points for early stage startups.
That being said - for DevOps people with experience across AWS/DO/Linode and Azure, this would be a good time to start the cold calls and sell your services to YC companies. Paying someone $5-10k on high ($2k/month) AWS bills to free services is a no-brainer.
EDIT: On that note, if any DevOps people want me to do the hustling (aka sales, lead generation) for them, let me know.
EDIT2: Anyone know if the $500k goes for 3-years? If it's only 1-year, then this may not beat the 1-2 years and ~$10k you get from Amazon via their platform: http://aws.amazon.com/activate/portfolio-detail/
Our new ceo force migrated us to office365 from gmail and our experience so far is not par with gmail. Very lacking or even unusable ui, not working or frustrating filters, very bad experience with imap clients unread email appearing read. And the very best part; there is no select all :)
And the shameless plug: Also your idea about selling migration services is very interesting. If any one is in need of system administration, integration, performance analysis, optimization, configuration management and devops like services I'm here to help.
They have two separate admin interfaces, neither of which you can do everything in so you have to break out powershell to finish off the job. The client had to lean heavily on me to help, something other clients have never had to do with gmail.
I like a lot of MS products, but from my brief exposure, exchange is a complete mess of WTFs. And annoyingly, googling doesn't often bring up pertinent answers and their help is pathetic.
Gmail is not flaky the way O365 is. Random outages with no explanation/contact from support until days after remedy...are no fun.
Otherwise, I don't see why small shops wanna move away from gmail to O365.
AWS and Azure are great, but when things get complicated, it means having good resources to help is critical.
note - from what I know, most early stage YC companies just use Heroku to solve many of these problems anyway.
If you're just running instances then maybe it's not too much work - though you're still going to have to think about backup / restore, access control, autoscaling (maybe), cost tracking / allocation, etc...
I'm sure Microsoft are banking on a profitable percentage of YC companies becoming so wedded to that Azure systems that the opportunity cost of switching to AWS is too high.
Edit: Typo
You can spin up your own load balancer pretty easily. DO offers backup too.
In terms of bang for the buck, you can't really beat DO, unless you're giving your services away for free ...
The post mentions that this is only for new YC startups, i.e. folks that won't have anything (or very little) to migrate.
Bundle that with cleaverly designed vendor lockin and General poor UX and you've got yourself a real headache.
We think this is a great way for startups to use the tons of cloud credits they get, instead of letting them go to waste or spending precious time on setting up infrastructure.
It's also worth pointing out that our $60k offer for YC companies is still around, if you're in one of the older batches.
Nietzsche: God is dead.
God (couple of years later): Nietzsche is dead.
pg: Microsoft is dead. http://www.paulgraham.com/microsoft.html
Microsoft (couple of years later): Here's some gift* for your startups. :)
(*: please don't start explaining that it is not a gift, etc. -- I know. :) )
Microsoft in the last 1-2 years is like a new scrappy competitor who lost market share on hubris. They are now willing to work how developers like to work not forcing them into the corral. Azure is the new OS at Microsoft, lock in higher up the stack but actually not as lock-in as before.
Microsoft went back to developer focus over biz/marketing focus and it is working like it always has.
=> As a Microsoft employee, i totally disagree. I would even go as far as guessing that you either are talking about something you don't know, or are just one of those evangelists that are paid to write such comments everywhere on the web.
Microsoft is a desperately consumer/business-oriented company, with no interest in technology whatsoever. If you don't get that, then you haven't been paying attention
- Buy all the good "Web 2.0" startups. They could get substantially all of them for less than they'd have to pay for Facebook.
- Put them all in a building in Silicon Valley, surrounded by lead shielding to protect them from any contact with Redmond.
I feel safe suggesting this, because they'd never do it.
I can't help but think that Microsoft has had quite a few acquisitions as of late, though. Minecraft (gaming), Acompli (mobile email), Sunrise (mobile calendar), and probably others.
They're a great company. Certly took advantage of the offer and their support and services have been excellent.
I was accepted into Microsoft's BizSpark program last year so I get $150/month free Azure credit. Azure is very nice to use. So far, I have just been using Ubuntu Linux VPS instances (similar to AWS EC2) and everything works similarly to other providers I use like AWS, IBM BlueMix, Digital Ocean, and Google's cloud services. Really, these are all very good services, so use the one you can get for free or at a low cost.
I used to be pretty much down on Microsoft, but now they are doing a lot of stuff right in my opinion: providing solid Linux support on Azure and providing an unbelievably good deal on the family edition of Office365. It seems like they are buying themselves into the cloud market, but with all of their cash on hand, that seems like a good strategy!
Anyway, if you have a business idea you want to try then look into BizSpark.
EDIT: correction, it looks like it's actually 5 users: http://blogs.msdn.com/b/niallsblog/archive/2014/08/22/activa...
I'm probably worried over nothing since YC investors are smart and tend to leave tech choices to the founders but I'd like to hear some feedback on this from people that pitch in the future.
Either way Microsoft seems to make pretty good decisions lately (from my humble POV). Seems like they are essentially playing the same power law as (other) investors, hoping that the next XXX will be run on Azure.
On a somewhat realted note...I worked in IT in Africa for a bit and it was pretty strange how locked into Microsoft server stuff they (Africa is big but it was true for all countires I had work relations in) are. I naively assumed that Linux/BSD would be more wide spread there but my very subjective conclusion is that MSxx-certificates are really widespread there and piracy is kind of tolerated so it's easy to learn. It was a little bit harder than expected to find Linux admins for example. I've always wondered if this was active corporate policy or just an "accident".
We were invited up to Redmond in the past to meet with teams from various Azure services and participate in private previews of some cool features. They're really working hard to make a great platform.
Satya, please don't have them mess things up with the next portal release! The whole sideways scrolling gives me a headache.
In the BizSpark Plus days, your company would receive $60K of hosting but could only use it to pay 100% of your bill in the first year of BS+. In the second year, you could use it to pay 50% and then in the third year you were paying full price for Azure. It was a pretty good deal, but if you are already spending $30-60K (or $250-$500 K) on infra in your first year or two, you're in a select breed of companies.
3 Years
1.) A working website, at least a landing page explaining your product/idea.
2.) An email address @ the same domain as the aforementioned website
If you have those two things you have a pretty solid shot at getting approved.
With a move like this Microsoft becomes THE Linux company as the millions of non-tech and tech businesses who are Windows based migrate to Quantum over time. They could make the underlying souped-up Linux OS FOSS and sell the Quantum OS layer (or whatever works).
I am thinking in terms of the next 50 years, not 5. In the long term I just don't see how the existing MS model can prevail. A move like this Quantum OS concept would instantly align them with web development and provide a solid alternative to Apple on developers desktops and perhaps even Ubuntu at the server level.
Passing them on at a 20% discount would make for a sweet ~$400k runway extension.
For the future YC batches - depending on what your company is going to be doing, it would make very little financial sense to NOT host on Azure.
However...
For a startup to choose to build on Azure they'll need staff with MSFT stack experience. 5 or 6 of those will burn through $500k in no time compared to the equivalent on a more open platform.
Whether or not to use Azure is certainly NOT obvious.
You can spin up bog standard Linux instances if you want. http://azure.microsoft.com/en-us/documentation/articles/virt...
You don't have to use a MSFT stack.
How do you figure that? Pricing on cloud services can be tough to compare so I really question anyone who presents blanket statements like that.
At the very least it will be interesting to see if more people adopt the Azure Websites PaaS offering. I know that right now most people dip their toes into Azure with IaaS because its historically more comfortable for them.
Why's that? Azure's just another cloud hosting provider. You don't have to run any kind of Microsoft software on the servers.
http://azure.microsoft.com/en-us/documentation/articles/virt...
edit: looks like phonon below beat me to it by a few minutes.
I dont have much experience with MS servers, as I like to get things done, but how are they when it comes to stability?
I had a friend who worked as a contractor for one of Microsoft's data centers a few years ago, and the stories he told me about how they run their datacenters was frightening. It sounded like management was out of touch and didn't really have a grasp on what was needed to keep everything running well.
Those stories leave me questioning Microsoft's basic competence in the cloud. While I'm glad to hear that they're improving their service, I just don't see them as being in the same class as AWS, Google, or even some of their smaller competitors such as DO. I would be wary of depending on Azure for anything critical.
If you'd like to share the specific stories this contractor told you, I'd be happy to talk you through it to figure out where your information went wrong.
Again, this was a couple of years ago. Also, I attended an event that MS held to show off Azure, and a large percentage of participants couldn't spin up an instance within 10 minutes, some of them didn't even start up. The MS people running the event had a direct line to some techs in the datacenter, but they still couldn't solve the problems. I talked with some MS partners/users who told me that Azure was terrible in their experience and even some MS execs admitted as much in private conversation. It sounds like Azure has improved since then, but I really have to wonder how much improvement has occured.
- Azure Cloud Services (PaaS)
- Azure Websites (PaaS light)
Unfortunately, when I last looked at it two months ago, both of them were quite unattractive, at least for Python hosting.
First of all, they have no support for automatically installing python dependencies, i.e. no support for pip. If you want to install e.g. django-rest-framework, you have to install it locally and then manually copy the installed files to Azure (with FTP, if I remember correctly).
Secondly, they both run Python on a Windows Server running behind the IIS web server. Besides the fact that you will walk along the road less travelled (Django is overwhelmingly deployed on Linux), there is a very concrete limitation: if you want to install binary packages (e.g. PIL or Pillow), you have to manually compile them for Windows (and of course copy them manually to your Azure server). Yuck.
If you want to use Azure, go with their IaaS offering (Azure Virtual Machines) and stay away from their PaaS offerings.
Been playing with CoreOS though...
http://venturebeat.com/2014/01/02/digitalocean-fixes-securit...
If you treat DO as just another VPS provider and pay on a monthly basis does it offer advantages over the competition?
Honestly, aceperry, I'm not seeing where any of this reflects poorly on our ability to operate x million machines, and the fact that you present it as evidence that we are not "competent" at provision our cluster effectively for our customers is worrying. I am starting to wonder if you know anything at all about the trade, because literally nothing here is relevant to your case. Even the hearsay you have about the DC vendors being incompetent is not really useful here, because it does not impact our operational capacity at all.
What am I missing in your statement that should cause me to worry?
In short, unless you have inside information, there is not enough information to determine which company is more competent or more trustworthy - it's not possible to judge which company you should trust with your business based on the available facts in an objective way currently, as the available data is easily attributable to simple bad luck.
In contrast, the performance difference is objectively measurable: http://joshtronic.com/2014/06/22/ten-dollar-showdown-linode-...
Regardless, I don't see how you can judge someone who's had at least two breaches more competent than someone who had zero. In all probability, DO didn't have any breaches precisely because they are more competent.